Fortunately, If you do get sued by the company you develop the OSS for, sourceXchange is not liable for any sum greater than the amount you get paid by them. So you may get taken to the cleaners, but they won't. Makes me feel all warm and fuzzy. --Shoeboy
Getting a degree also represents that you are willing to work hard for something and you can finish it, that you actually know how to work within a system to accomplish something. Yes, and joining a cult indicates that you are good at following orders, respectful of authority, willing to put the good of others before your own... Getting a degree _REALLY_ means that you're willing to spend 18 Grand of your parents money learning something that you could have picked up much more quickly and cheaply by taking a little initiative. If you really wan't to know about the Napoleanic wars, the best method is to read history books, biographies of the major players and contemporary letters and journals. You can learn more in less time than you would spending 5hrs a week in lecture and 10Hrs a week doing homework. Oh and by the way, the team I work on is second to none - and my fellows would happilly agree that I can run with the best of them. --PJ
So now we have 4 banks of equally valid statistics (well not really, since the online, fact. install and piracy stats are best guesses) The neato thing is that we can look at the pirated software category and say that MS is kicking butt. Linux has 0% market share. That means it's not desirable (actually it means it's free) We can look at the online downloads and see that MS market share is effectively zero since only a few warez d00dz are downloading it. We can look at factory installs and assume that MS has a healthy lead and that competition is limited. None of these stats is false, but none of them gives a true representation of software usage patterns in Japan. I personally have no idea what % of OS's in Japan are sold via retail channels. This stat suggests that given an array of competing products lined up on a shelf the Japanese tend to buy Linux at the same rate as Win98. What it does not suggest is that MS market share in Japan is endangered. It might be, but we can't tell this by looking at a highly specific branch of the OS trade. --Shoeboy
This one's misleading since it breaks Win98 sales into 4 parts. Added together you get: 29.9 All Win98 10.23 Mac 28.76 All linux This is really amazing for the linux crowd, and espescially turbo linux, but it isn't the victory y'all been waiting for. Just a further narrowing of the gap. It should however, be taken as a victory for all of you that hate redhat for some odd reason or other. --Shoeboy.
If you truly are a high-school dropout What? Of course I'm a high school dropout. Didn't you notice the creative way I spelled fascism?;-) Seriously though, the assertion that completing college guarantees superior abilities is a little weak. I went to college for a year. (yes, they do let h.s. dropouts go to college. Most college admissions programs look at your current GPA + SAT/ACT only, there is no point value to graduation in their system.) While in college I learned an awful lot. I learned how to get drinks despite being underage. I learned that If you get a woman drunk enough she'll mistake you for Brad Pitt. I learned that if you get drunk enough, you'll start mistaking most women (and a few men) for Claudia Schiffer. I learned how to roll a joint. The lessons were endless. I'm not sure that sticking with college for 4 years would have made me a more effective worker. I'm pretty sure that I would have experienced VD, liver damage and jail time. That's why I get confused when people talk about the benefits of a degree. --Shoeboy
Right on brother. I don't like it when people I can outperform with one frontal lobe tied behind my back get paid more than I do because they are 35yr old college grads and I'm a 22yr old high school dropout. Unions would tend to enforce a salary scale that makes that happen. That's bad. People should be able to band together to bargain with their employers. This is true. The argument that people should ALWAYS band together and make demands is ridiculous. This line of reasoning benefits the managers of the unions at the expense of the workers. In short, those workers who cling blindly to the union idea are encouraging the union to exploit them instead of allowing the capitalists to exploit them. Sort of like exchanging facism for stalinism. Nice.
Ok, guilty confession time. I used to work on the main MS campus. It sucked. I made $19.38/hr doing the same work that full time MS employees made 50K + stock doing, and other temps made $30-$40/hr. So I was one of them poor exploited underpaid temps. But wait... I was 20 years old and making more $$$ than my father ever did. That's pretty good. And it took me a week to make as much $$$ as the average hatian makes in a year. Even better. Then I quit and got a job with a different company for lots more money. Funny, I don't seem to be able to convince anyone that I was exploited. --Shoeboy
It's evolution in action. What I long for is the day that we get "And Sometimes 'Y'" prefixing everything. I'm there already. Sometimes I look at my blue screen and say "yMicrosoft."
Oracle on NT is really amazingly stupid. Oracle was designed for Unix. Unix and NT have radically different process models (CreateProcess() vs. fork() ) and the NT model has 3x the process creation overhead. Oracle would have to be totally rewritten to run efficiently on NT, and it hasn't been. MSSQL runs rings around it since MSSQL has been designed to run on NT. If you want to run Oracle on NT, you obviously don't care about performance, and if you don't care about performance, why are you buying a $12k piece of software? The moral of all this? Run oracle on unix. Specifically, run oracle for Tru64 Unix on a Compaq GS-140! --Shoeboy
It's doubled approximately 28 times to get to where it is now, and at a very predictable rate. It only needs to double 4 more times to reach 60%. Given that it's doubled 28 times without much trouble, well.. you do the math. OK, I did the math. We start with 1 user. We double that 28 times. We now have have 2^28 users or ~268Million users. We double 4 more times, that's 2^32 users. ~4Billion users. That's not 60% of the server market, that's 60% of the entire planet earth. We double once more and we have 2^33 users. Wow. We're now out of people to use linux. We'll have to start marketing to the aliens that transmeta keeps in its underground bunkers.
Running in a 12-way configuration the system delivered a SPECweb96 benchmark of 40,161 http ops/sec, besting Hewlett-Packard's HP 9000 N-Class server by 66 percent. This is pathetic - It would take this box over 112 days to service an http request from every man, woman and child on the planet earth. This just isn't going to cut it. I need a machine that can server a number of pages/day greater than the population of France. This shabby IBM box falls short by almost a million. Seriously, If you use one of those beauties to serve static web pages, you need your head examined. It's like using the USS Missouri for island hopping. I'd like to see TPC/H numbers though - data warehousing is one of the few apps that could use this kind of horsepower. --Shoeboy
At first I was upset by this story. I thought they said beaver.com had had its web access revoked. For a while there I thought the calluses on my palm were going to fade.
BuyThisDomainYouMoron.com OverPricedDomainName.com YouMustBeAnIdiotIfYouBuyThis.com And only $65K per domain. Why hasn't anyone bought these? Any takers? --Shoeboy
If they're going to network cars, the first thing we need is a better collision avoidance algoritm. The standard ethernet method is ok: if your car is in a collision you wait a random amount of time (determined by the speed of your mechanic) and try again. The problem with this is the high fatality rate. Token ring can't work, because we'd have to replace all our stop lights with 4 way stops. This will be more expensive and kill average throughput. Regardless of how they do this, I don't want to be considered an acceptable packet loss statistic. --Shoeboy
I work with a guy who claims to have an excellent fed spotting tactic. The feds all have pristine laptops - govmnt property after all, but lotsa hackers plaster stickers and crap all over their laptops. Stickers == NO_FED
If you read the original post, I asked for meaningless statistics of dubious origin. (or something like that) I was operating on 3 hours sleep and feeling cranky - my original intent was to gripe about both the constant, completely unsupported 'open source works better' assertions, and the amazing prevalence of lousy numbers on the web. I didn't do it coherently though, sorry. Fortunately my request was granted, and I got lots of meaningless statistics of dubious origin. I'm pretty sure that Alan isn't a dubious source for linux development info, but I am pretty sure all of the statistics in the thread lack meaning - more so, since we have no idea how they were gathered (except my 35MLOC number, and that was a wild guess. Other guesses I've seen about the web number it from 12MLOC to 65MLOC) and precious little decent analysis of what they mean. --Shoeboy
If debugged lines of code were nose rings, in 33years and 4 months, the linux hackers could decorate a nostril on every Algerian woman between the ages of 15 and 65. (based on July 1998 estimate.) Furthermore, if the linux hackers recieved $100 per debugged LOC, in 4581 years they would be able to make 120.4 billion purchases at the dollar store.
Oh yeah, and you should start by assuming that the 35M LOC in windows is a best guess. You should follow it up by assuming that Alan's.5LOC/min number cannot possibly indicate either further kernel development rates, or the rate of accelaration in development rates over time - there's no way to measure this without a time machine. It would also be worth noting that MS code is in a large part wizard generated, and may not be strictly neccessary. While were at it, we can note that #include "stdio.h" #include "tchar.h" #include "windows.h" #define HELLO_MSG "Hello World.\n" void main(void); void main() { _tprintf(_T(HELLO_MSG)); }
Lets have some fun. Lets assume that there is one bug per every 50 lines of windows code. This means that a team writing 1 line of buggy code every 2 minutes (lets call them the lobotomized linux hackers) would take ~2.6 years just to = the number of bugs in windows. For mor fun, we take the.5Loc/Min number and times it by 8 years - the time that linux has been around - this means that the linux kernel is just over 2 million LOC. Is this true? Now, lets assume that the linux hackers got tired of linux, and are writing #include void main(void) { printf("Hello World!\n"); } over and over again. They could produce 87600 versions of hello world in a single year. Yeah. Now lets see how MS stacks up - NT has been in development for 10 years (a guess, and a nice round number), it has 35000000 LOC (another guess) This has them writing about 6.5 LOC/Minute. They could write over a million copies of "hello world" per year, except the MFC version of hello world is huge, so no, they really couldn't. I'm way off topic now, and I don't know what any of this has to do with anything. --Shoeboy
1 LOC per 2 minutes = 720 LOC/day. Windows 2000 ~= 35000000 LOC Unless I hit the keys on my calculator wrong, this means that it would take the Linux hackers ~133 years to write W2K. Wonder how that compares to MS? (insert your own W2K shipping delay joke here) No wonder the free software community is opposed to bloat. --Shoeboy
Ok, I keep seeing posts claiming that MS is paying people to post on/. If this is true, it totally redeems them in my view. In fact, I can't think of a better job. Free soda and/. Where do I apply? Do you get bonuses for "First Posts?" The one downside would be the performance reviews - "Well shoeboy, I'd like to give you a raise, but you keep getting moderated down for being off topic." --Shoeboy
Open source enthusiasts need to put numbers behind their claims. I'd like to see a list of say KLOC's per second of patches generated by open source. Or maybe a user happiness quotient. These essays are good and all, but how can we make decisions without dubious statistics gathered by unspecified methods?
Slashdot Mirror
Fortunately, If you do get sued by the company you develop the OSS for, sourceXchange is not liable for any sum greater than the amount you get paid by them. So you may get taken to the cleaners, but they won't. Makes me feel all warm and fuzzy.
--Shoeboy
Yeah, cert is way easy to get. It took me 6weeks to get my MCSE and I had never even seen NT.
--Shoeboy
Getting a degree also represents that you are willing to work hard for something and you can finish it, that you actually know how to work within a system to accomplish something.
Yes, and joining a cult indicates that you are good at following orders, respectful of authority, willing to put the good of others before your own...
Getting a degree _REALLY_ means that you're willing to spend 18 Grand of your parents money learning something that you could have picked up much more quickly and cheaply by taking a little initiative. If you really wan't to know about the Napoleanic wars, the best method is to read history books, biographies of the major players and contemporary letters and journals. You can learn more in less time than you would spending 5hrs a week in lecture and 10Hrs a week doing homework.
Oh and by the way, the team I work on is second to none - and my fellows would happilly agree that I can run with the best of them.
--PJ
I'm willing to bet that the breakdown in japan goes something like this:
... 2%
Retail Outlets:
Win98 30%
MacOS 10%
Linux 28%
InternetDownloads:
Linux 80%
*BSD 18%
PreInstalled versions
Win9x 80%
WinNT 7%
MacOS 11%
Linux/*Nix/BEOS/CPM/etc
Pirated Software
Win9x 90%
WinNT 5%
Proprietary *Nix 2%
NetWare 3%
So now we have 4 banks of equally valid statistics (well not really, since the online, fact. install and piracy stats are best guesses) The neato thing is that we can look at the pirated software category and say that MS is kicking butt. Linux has 0% market share. That means it's not desirable (actually it means it's free) We can look at the online downloads and see that MS market share is effectively zero since only a few warez d00dz are downloading it. We can look at factory installs and assume that MS has a healthy lead and that competition is limited. None of these stats is false, but none of them gives a true representation of software usage patterns in Japan. I personally have no idea what % of OS's in Japan are sold via retail channels. This stat suggests that given an array of competing products lined up on a shelf the Japanese tend to buy Linux at the same rate as Win98. What it does not suggest is that MS market share in Japan is endangered. It might be, but we can't tell this by looking at a highly specific branch of the OS trade.
--Shoeboy
This one's misleading since it breaks Win98 sales into 4 parts. Added together you get:
29.9 All Win98
10.23 Mac
28.76 All linux
This is really amazing for the linux crowd, and espescially turbo linux, but it isn't the victory y'all been waiting for. Just a further narrowing of the gap. It should however, be taken as a victory for all of you that hate redhat for some odd reason or other.
--Shoeboy.
This is a quote from scott mcnealy, right? Or is it Kant?
--Shoeboy
If you truly are a high-school dropout ;-)
What? Of course I'm a high school dropout. Didn't you notice the creative way I spelled fascism?
Seriously though, the assertion that completing college guarantees superior abilities is a little weak. I went to college for a year. (yes, they do let h.s. dropouts go to college. Most college admissions programs look at your current GPA + SAT/ACT only, there is no point value to graduation in their system.) While in college I learned an awful lot. I learned how to get drinks despite being underage. I learned that If you get a woman drunk enough she'll mistake you for Brad Pitt. I learned that if you get drunk enough, you'll start mistaking most women (and a few men) for Claudia Schiffer. I learned how to roll a joint. The lessons were endless. I'm not sure that sticking with college for 4 years would have made me a more effective worker. I'm pretty sure that I would have experienced VD, liver damage and jail time. That's why I get confused when people talk about the benefits of a degree.
--Shoeboy
Apologies for the spelling. It's early and I'm tired.
Right on brother. I don't like it when people I can outperform with one frontal lobe tied behind my back get paid more than I do because they are 35yr old college grads and I'm a 22yr old high school dropout. Unions would tend to enforce a salary scale that makes that happen. That's bad.
People should be able to band together to bargain with their employers. This is true. The argument that people should ALWAYS band together and make demands is ridiculous. This line of reasoning benefits the managers of the unions at the expense of the workers. In short, those workers who cling blindly to the union idea are encouraging the union to exploit them instead of allowing the capitalists to exploit them. Sort of like exchanging facism for stalinism. Nice.
Ok, guilty confession time. I used to work on the main MS campus. It sucked. I made $19.38/hr doing the same work that full time MS employees made 50K + stock doing, and other temps made $30-$40/hr. So I was one of them poor exploited underpaid temps. But wait... I was 20 years old and making more $$$ than my father ever did. That's pretty good. And it took me a week to make as much $$$ as the average hatian makes in a year. Even better. Then I quit and got a job with a different company for lots more money. Funny, I don't seem to be able to convince anyone that I was exploited.
--Shoeboy
It's evolution in action. What I long for is the day that we get "And Sometimes 'Y'" prefixing everything. I'm there already. Sometimes I look at my blue screen and say "yMicrosoft."
Oracle on NT is really amazingly stupid. Oracle was designed for Unix. Unix and NT have radically different process models (CreateProcess() vs. fork() ) and the NT model has 3x the process creation overhead. Oracle would have to be totally rewritten to run efficiently on NT, and it hasn't been. MSSQL runs rings around it since MSSQL has been designed to run on NT. If you want to run Oracle on NT, you obviously don't care about performance, and if you don't care about performance, why are you buying a $12k piece of software? The moral of all this? Run oracle on unix. Specifically, run oracle for Tru64 Unix on a Compaq GS-140!
--Shoeboy
It's doubled approximately 28 times to get to where it is now, and at a very predictable rate. It only needs to double 4 more times to reach 60%. Given that it's doubled 28 times without much trouble, well.. you do the math.
OK, I did the math. We start with 1 user. We double that 28 times. We now have have 2^28 users or ~268Million users. We double 4 more times, that's 2^32 users. ~4Billion users. That's not 60% of the server market, that's 60% of the entire planet earth. We double once more and we have 2^33 users. Wow. We're now out of people to use linux. We'll have to start marketing to the aliens that transmeta keeps in its underground bunkers.
Running in a 12-way configuration the system delivered a SPECweb96 benchmark of 40,161 http ops/sec, besting Hewlett-Packard's HP 9000 N-Class server by 66 percent.
This is pathetic - It would take this box over 112 days to service an http request from every man, woman and child on the planet earth. This just isn't going to cut it. I need a machine that can server a number of pages/day greater than the population of France. This shabby IBM box falls short by almost a million.
Seriously, If you use one of those beauties to serve static web pages, you need your head examined. It's like using the USS Missouri for island hopping. I'd like to see TPC/H numbers though - data warehousing is one of the few apps that could use this kind of horsepower.
--Shoeboy
At first I was upset by this story. I thought they said beaver.com had had its web access revoked. For a while there I thought the calluses on my palm were going to fade.
BuyThisDomainYouMoron.com
OverPricedDomainName.com
YouMustBeAnIdiotIfYouBuyThis.com
And only $65K per domain. Why hasn't anyone bought these? Any takers?
--Shoeboy
If they're going to network cars, the first thing we need is a better collision avoidance algoritm.
The standard ethernet method is ok: if your car is in a collision you wait a random amount of time (determined by the speed of your mechanic) and try again. The problem with this is the high fatality rate.
Token ring can't work, because we'd have to replace all our stop lights with 4 way stops. This will be more expensive and kill average throughput.
Regardless of how they do this, I don't want to be considered an acceptable packet loss statistic.
--Shoeboy
I work with a guy who claims to have an excellent fed spotting tactic. The feds all have pristine laptops - govmnt property after all, but lotsa hackers plaster stickers and crap all over their laptops. Stickers == NO_FED
If you read the original post, I asked for meaningless statistics of dubious origin. (or something like that) I was operating on 3 hours sleep and feeling cranky - my original intent was to gripe about both the constant, completely unsupported 'open source works better' assertions, and the amazing prevalence of lousy numbers on the web. I didn't do it coherently though, sorry. Fortunately my request was granted, and I got lots of meaningless statistics of dubious origin. I'm pretty sure that Alan isn't a dubious source for linux development info, but I am pretty sure all of the statistics in the thread lack meaning - more so, since we have no idea how they were gathered (except my 35MLOC number, and that was a wild guess. Other guesses I've seen about the web number it from 12MLOC to 65MLOC) and precious little decent analysis of what they mean.
--Shoeboy
If debugged lines of code were nose rings, in 33years and 4 months, the linux hackers could decorate a nostril on every Algerian woman between the ages of 15 and 65. (based on July 1998 estimate.) Furthermore, if the linux hackers recieved $100 per debugged LOC, in 4581 years they would be able to make 120.4 billion purchases at the dollar store.
.5LOC/min number cannot possibly indicate either further kernel development rates, or the rate of accelaration in development rates over time - there's no way to measure this without a time machine. It would also be worth noting that MS code is in a large part wizard generated, and may not be strictly neccessary. While were at it, we can note that
Oh yeah, and you should start by assuming that the 35M LOC in windows is a best guess.
You should follow it up by assuming that Alan's
#include "stdio.h"
#include "tchar.h"
#include "windows.h"
#define HELLO_MSG "Hello World.\n"
void main(void);
void main() {
_tprintf(_T(HELLO_MSG));
}
is equivalent to:
#include "stdio.h"
void main(void) { printf("Hello World.\n");}
even though 1 has many more LOC. In conclusion, 75% of all statistics are meaningless.
--Shoeboy
Lets have some fun. Lets assume that there is one bug per every 50 lines of windows code. This means that a team writing 1 line of buggy code every 2 minutes (lets call them the lobotomized linux hackers) would take ~2.6 years just to = the number of bugs in windows. For mor fun, we take the .5Loc/Min number and times it by 8 years - the time that linux has been around - this means that the linux kernel is just over 2 million LOC. Is this true? Now, lets assume that the linux hackers got tired of linux, and are writing
#include
void main(void) {
printf("Hello World!\n"); }
over and over again. They could produce 87600 versions of hello world in a single year. Yeah. Now lets see how MS stacks up - NT has been in development for 10 years (a guess, and a nice round number), it has 35000000 LOC (another guess) This has them writing about 6.5 LOC/Minute. They could write over a million copies of "hello world" per year, except the MFC version of hello world is huge, so no, they really couldn't. I'm way off topic now, and I don't know what any of this has to do with anything.
--Shoeboy
1 LOC per 2 minutes = 720 LOC/day.
Windows 2000 ~= 35000000 LOC
Unless I hit the keys on my calculator wrong, this means that it would take the Linux hackers ~133 years to write W2K. Wonder how that compares to MS? (insert your own W2K shipping delay joke here) No wonder the free software community is opposed to bloat.
--Shoeboy
Ok, I keep seeing posts claiming that MS is paying people to post on /. If this is true, it totally redeems them in my view. In fact, I can't think of a better job. Free soda and /. Where do I apply? Do you get bonuses for "First Posts?" The one downside would be the performance reviews - "Well shoeboy, I'd like to give you a raise, but you keep getting moderated down for being off topic."
--Shoeboy
Open source enthusiasts need to put numbers behind their claims. I'd like to see a list of say KLOC's per second of patches generated by open source. Or maybe a user happiness quotient. These essays are good and all, but how can we make decisions without dubious statistics gathered by unspecified methods?
Kernel32.dll is ~ 700Kb in size. Hal.dll for x86 is ~90Kb. On W2K beta 3 at any rate.
--Shoeboy